Changing Tilt ?

[Ken82]

It became apparent to me last night that I had some sort of changing tilt in my images which was limiting the ability of star alignment. 
 

The following short video shows one evening of imaging (no meridian flip) and a few times refocusing in between. I’m measuring FWHM to get these graphs. 

The first images are fairly concentric with the best FWHM in the centre as expected. As the night moves on the best FWHM area clearly moves down and off centre and then into the bottom left hand corner. The poor area of focus starts to creep into the top right hand corner. 
 

For these images all auto focusing was done on centre. 
 

so firstly I’m not sure FWHM is a good method to test ? 
 

In terms of the difference in FWHM between best and worse 1.5-3 FWHM can this give us an indicator how large the tilt could be ? It may even be fairly good values and doesn’t show much tilt, but I’m not sure what I’m looking at here. I would expect the values to be concentric? 
 

What could be tilting ? I would expect the sensor to have at least a small amount of tilt but this shouldn’t change during the night ?

Adapters are another likely cause but camera and filter wheel are both bolted. The only other connection is m72 to m68 (It’s fairly robust though ?!)

Or even focuser but again this is fairly robust !?
 

I may have been looking in the wrong place to resolve my other problems. Maybe it is tilt I need to correct which will solve my other issue 😀
 

Don’t know where to start if I’m honest 😬

[michael8554]

Assuming that your video covered a wide range of sky and so different camera orientations. looks like there is a fixed tilt, not image train droop.

If in doubt take an image, then another of the same target, with the camera rotated 90 degrees.

So one of these may help:

https://www.firstlightoptics.com/adapters/astro-essentials-m42-t-tilt-adjuster.html

Michael

[Ken82]

5 hours ago, michael8554 said:

Assuming that your video covered a wide range of sky and so different camera orientations. looks like there is a fixed tilt, not image train droop.

If in doubt take an image, then another of the same target, with the camera rotated 90 degrees.

So one of these may help:

https://www.firstlightoptics.com/adapters/astro-essentials-m42-t-tilt-adjuster.html

Michael

Thanks Michael you could be right I’m way over my head here. If it was fixed tilt why does it start off concentric though and slowly move to the bottom right ? 
 

is measuring FWHM a reasonable way to test tilt ? 
 

ken 

[michael8554]

Perhaps you're right, my quick look suggested the error was diagonal.

If everything is screwed then that only leaves the focuser.

I know from modding Canon DSLR's that they adjust the sensor tilt to 0.03mm tolerance.

That's not a lot of movement !

What kit are we talking about, you haven't said ?

Michael

[Ken82]

2 hours ago, michael8554 said:

Perhaps you're right, my quick look suggested the error was diagonal.

If everything is screwed then that only leaves the focuser.

I know from modding Canon DSLR's that they adjust the sensor tilt to 0.03mm tolerance.

That's not a lot of movement !

What kit are we talking about, you haven't said ?

Michael

Yea I’m not sure really maybe someone else can give a viewpoint on the quick video ? 
 

The focuser is the new 4” one on the fsq edxiv so pretty solid although it does need an auto focuser to stop slippage. Looking at it I really can’t see how this could sag. 
 

Since it starts of concentric and slowly moves over the evening this is the only reason I thought it was moving tilt.  It could be seeing conditions ? No idea what else  😕
 

I’m not sure a difference of 1.5 FWHM on a large field is bad anyway ?? My thinking was that it should at least be concentric with the sharpest stars in the middle ? 
 

Edit- I’m in Germany now for a week so I’ll have a look when home. Maybe have to quarantine 🥳

Edited August 27, 2020 by Ken82

[Ken82]

Just bought myself a m68 tilt adapter 

https://www.365astronomy.com/camera-tilting-unit-xl-ctu-m68x1.html

Hopefully it’s in the post box when I get back from Germany. 
 

Just a thought I’ve been measuring FWHM up to now. Can eccentricity tell us anything ?

Image on left is this scope and right hand side is another scope With same camera etc. I think they are both good ?? But should they both not be concentric??

[Ken82]

In between my slipping focuser I had a look for tilt the last couple of nights in my subs. 
 

From close inspection of the corners I think the field is very flat but a slight tilt to the upper right hand corner which would support the FWHM theory. 
 

I think the stars in all four corners show a very mild sagittal astigmatism but the slightly longer elongation of the upper right hand corner is showing tilt. 
 

I think it’s excellent for such a large sensor but then I’m no expert. 
 

Maybe just a slight adjustment of the m68 tilt adjuster ?? 
 

Here is one sub if anyone wants to look -

https://drive.google.com/file/d/1RigF6f_sKsnnByFi8WCiqNol94ufDGWI/view?usp=drivesdk
 

ken 

[Ken82]

My Gert Neumann M68 tilt adapter arrived this morning. It looks a real robust piece of kit. 
 

Unfortunately I’ve discovered since ordering this that my focuser is unable to hold the weight of the camera without slipping. So adding a further 350g is not really an option. This would take the weight up to 2.5kg. 
 

I’ll see what Ian thinks about the focuser on Monday and then decide what to do. 
 

For now I’ll put it together on my other scope and see if i can adjust out the remaining tilt. 
 

ken 

[symmetal]

Hi Ken82,

I ran your last posted image through CCD Inspector and it does report that the main issue is tilt as shown on the following charts: The 3D FWHM chart shows the consistant change from bottom left to top right. The aspect chart is very similar to your eccentricity plot

My ASI6200 full frame camera had tilt and I found that SGPs auto-focus always focused on the top left. Manually altering the focus by 5 or 10 steps would alter the best focus point to different areas of the frame including the centre. I bought the 48mm Gert Neumann tilt unit. The 68mm one you have was a lot more expensive. They are much easier to use compared to the cheaper units with the push-pull screws. For the price they should be too. 😀 I had to get the FF spacing fairly close before the tilt adjuster had much effect. As yours is a quad you don't have that problem. You can try to work out which screw to adjust first but it's easier to give one a full turn or so, and see what affect it has. CCDI 3D plot would give me a quick assessment of the affect. Your FWHM plots can be interpreted similarly I hope, but if you want to post another image to check, I can run it through CCDI for you. 🙂

I don't know what the percentage figures for curvature, and particularly tilt actually refer to, but it doesn't really matter as the plots tell the story. 😀

FWHM Curvature Plot

3D FWHM Curvature Plot

Star Aspect Chart

Alan

[Ken82]

17 hours ago, symmetal said:

Hi Ken82,

I ran your last posted image through CCD Inspector and it does report that the main issue is tilt as shown on the following charts: The 3D FWHM chart shows the consistant change from bottom left to top right. The aspect chart is very similar to your eccentricity plot

My ASI6200 full frame camera had tilt and I found that SGPs auto-focus always focused on the top left. Manually altering the focus by 5 or 10 steps would alter the best focus point to different areas of the frame including the centre. I bought the 48mm Gert Neumann tilt unit. The 68mm one you have was a lot more expensive. They are much easier to use compared to the cheaper units with the push-pull screws. For the price they should be too. 😀 I had to get the FF spacing fairly close before the tilt adjuster had much effect. As yours is a quad you don't have that problem. You can try to work out which screw to adjust first but it's easier to give one a full turn or so, and see what affect it has. CCDI 3D plot would give me a quick assessment of the affect. Your FWHM plots can be interpreted similarly I hope, but if you want to post another image to check, I can run it through CCDI for you. 🙂

I don't know what the percentage figures for curvature, and particularly tilt actually refer to, but it doesn't really matter as the plots tell the story. 😀

FWHM Curvature Plot

3D FWHM Curvature Plot

Star Aspect Chart

Alan

Thanks very much for your feedback Alan 👍

I suspect your right ,  the tilt is within The camera. I think the tilt adapter should solve this. 
 

I thought 4 adjustment points (one each corner) would have been easier/better but I’ll see how I get on. why only 3 though ??
 

I did have a quick look on Monday but it was rather cloudy and I didn’t get any decent subs. 
 

ive actually agreed to send the scope back now as the focuser kept slipping without fully locking so I couldn’t automate things (thanks to FLO for that BTW) . Maybe that was premature and I should have got the feathertouch 🤔 oh I don’t know anymore ! He says as he browses the feathertouch on 365 astronomy! 

I did note on Monday that I was able to get the star shapes good in the corners but then refocused and they were bad again. Not sure what’s going on there really but it could have even been the focuser slipping. 

 Thanks ken 

 

[symmetal]

1 hour ago, Ken82 said:

Thanks very much for your feedback Alan 👍

I suspect your right ,  the tilt is within The camera. I think the tilt adapter should solve this. 
 

I thought 4 adjustment points (one each corner) would have been easier/better but I’ll see how I get on. why only 3 though ??
 

I did have a quick look on Monday but it was rather cloudy and I didn’t get any decent subs. 
 

ive actually agreed to send the scope back now as the focuser kept slipping without fully locking so I couldn’t automate things (thanks to FLO for that BTW) . Maybe that was premature and I should have got the feathertouch 🤔 oh I don’t know anymore ! He says as he browses the feathertouch on 365 astronomy! 

I did note on Monday that I was able to get the star shapes good in the corners but then refocused and they were bad again. Not sure what’s going on there really but it could have even been the focuser slipping. 

 Thanks ken 

Yes the tilt adjuster should correct that. Depending on the cause of the tilt, (it may be more than one item in the image train causing it) you may have to keep the orientation of the camera in one position to avoid tilt re-occuring on rotating the camera. For 99% of the images I always have the image long axis parallel to RA anyway, so that isn't really a problem. Only the horsehead looks better with a 90 degree rotation. 😀

As the tilt adjuster just screws on it's a bit of luck where the adjustment screws lie with respect to the camera orientation so they most likely wouldn't be at the corners. Having three screws ensures they are always all pushing on the plate to keep it in one position. With four screws every time you adjusted one screw you would have to retighten one or two other screws to maintain the force on the plate. The lower cost push-pull screw type adjusters also generally need all the screws adjusted each time to maintain even force on the plate so are more fiddley.

Yes, large cameras/filter wheels and the tilt adjuster cause extra strain on the focuser. At least my rack and pinion one is fine in that respect. 🙂

Altering the focus and being able to get good stars in the corner implies that while your quad scope is correcting the corner star aberrations well, there is some residual field curvature left. You can't cure this with tilt adjustments but you should be able to get the focus aberrations to a minimum and matched at all four corners.

Auto-focus usually goes for the lowest average FWHM over the whole frame, possibly weighted towards the centre, so the corners tend to be worse off using auto-focus. I found manually focusing out by 5 to 10 steps after auto-focus will give a sharper image at the centre and possibly slightly worse corner stars. If the main subject is at the centre it can be better to do this.

Alan

[Ken82]

On 16/09/2020 at 21:31, symmetal said:

Yes the tilt adjuster should correct that. Depending on the cause of the tilt, (it may be more than one item in the image train causing it) you may have to keep the orientation of the camera in one position to avoid tilt re-occuring on rotating the camera. For 99% of the images I always have the image long axis parallel to RA anyway, so that isn't really a problem. Only the horsehead looks better with a 90 degree rotation. 😀

As the tilt adjuster just screws on it's a bit of luck where the adjustment screws lie with respect to the camera orientation so they most likely wouldn't be at the corners. Having three screws ensures they are always all pushing on the plate to keep it in one position. With four screws every time you adjusted one screw you would have to retighten one or two other screws to maintain the force on the plate. The lower cost push-pull screw type adjusters also generally need all the screws adjusted each time to maintain even force on the plate so are more fiddley.

Yes, large cameras/filter wheels and the tilt adjuster cause extra strain on the focuser. At least my rack and pinion one is fine in that respect. 🙂

Altering the focus and being able to get good stars in the corner implies that while your quad scope is correcting the corner star aberrations well, there is some residual field curvature left. You can't cure this with tilt adjustments but you should be able to get the focus aberrations to a minimum and matched at all four corners.

Auto-focus usually goes for the lowest average FWHM over the whole frame, possibly weighted towards the centre, so the corners tend to be worse off using auto-focus. I found manually focusing out by 5 to 10 steps after auto-focus will give a sharper image at the centre and possibly slightly worse corner stars. If the main subject is at the centre it can be better to do this.

Alan

Ive had a look through Mondays data from mapping the CTU and i think its now looking better? A little bit of residual astigmatism (as expected with 3.75um pixels) but the tilt adapter appears to have done a decent job without actually trying to correctly adjust the tilt. What do you think?

https://drive.google.com/file/d/1Xz0UWUfXacsWje1bYGYqXjmJb9e7gL1c/view?usp=sharing

Thanks ken 

[symmetal]

14 minutes ago, Ken82 said:

Ive had a look through Mondays data from mapping the CTU and i think its now looking better? A little bit of residual astigmatism (as expected with 3.75um pixels) but the tilt adapter appears to have done a decent job without actually trying to correctly adjust the tilt. What do you think?

https://drive.google.com/file/d/1Xz0UWUfXacsWje1bYGYqXjmJb9e7gL1c/view?usp=sharing

Thanks ken 

It looks like you have an automatic tilt adjuster there. Well worth the price. 😄 Here are the corresponding CCDI plots: Tilt is almost perfect. None in Y plane. Very slight in X, shown by the left edge being slightly higher than the right. It's reduced from 1.1" to 0.2". The min to max in FWHM has gone from 2.91 - 4.21 to 2.47 - 3.20 which is significant and best focus is now in the centre. The aspect has improved too although the colours may not suggest that. It seems to always plot min to max as black to pink. Aspect has gone from 6 - 76 to a much closer 8 - 37.

By coincidence, like my FLT-98, stars on the far left tend to be radial while those on the far right tend to be tangential with a hint of coma. Mine are a bit worse than yours in some respects, particularly the coma. I found altering FF spacing and tilt could make them worse but they always had the same type of distortion. It looks like with a sensor this size you maybe need to spend a lot more on the scope to improve it unless you're lucky. 🙂

The 3D plot just corresponds to star FWHM values and worse values plot higher in the Z plane. This indicates tilt if not focused in the centre like the previous image. You aim for the X and Y FWHM 'curvature' to be symmetrical about the centre which yours pretty much now are. 🙂

Curvature

3D

Aspect

Alan

[Ken82]

On 17/09/2020 at 23:47, symmetal said:

It looks like you have an automatic tilt adjuster there. Well worth the price. 😄 Here are the corresponding CCDI plots: Tilt is almost perfect. None in Y plane. Very slight in X, shown by the left edge being slightly higher than the right. It's reduced from 1.1" to 0.2". The min to max in FWHM has gone from 2.91 - 4.21 to 2.47 - 3.20 which is significant and best focus is now in the centre. The aspect has improved too although the colours may not suggest that. It seems to always plot min to max as black to pink. Aspect has gone from 6 - 76 to a much closer 8 - 37.

By coincidence, like my FLT-98, stars on the far left tend to be radial while those on the far right tend to be tangential with a hint of coma. Mine are a bit worse than yours in some respects, particularly the coma. I found altering FF spacing and tilt could make them worse but they always had the same type of distortion. It looks like with a sensor this size you maybe need to spend a lot more on the scope to improve it unless you're lucky. 🙂

The 3D plot just corresponds to star FWHM values and worse values plot higher in the Z plane. This indicates tilt if not focused in the centre like the previous image. You aim for the X and Y FWHM 'curvature' to be symmetrical about the centre which yours pretty much now are. 🙂

Curvature

3D

Aspect

Alan

Thanks Alan, I’ve had time to download ccdi on holiday and I can see that for myself. 
 

I’m pleased with this TBH. The tilt from adapters and sensor has been corrected quite easily, maybe I can even better refine this in the future. 
 

You mention perhaps having to spend a lot more money to improve it or getting lucky. Honestly I was under the impression I had already done this by spending 5-6k on a fsq. I’m actually still of the belief the fsq is likely the best scope for large sensors like we are using. But what is the alternative even if we spend countless more ? 
 

Thanks ken 

[symmetal]

3 hours ago, Ken82 said:

You mention perhaps having to spend a lot more money to improve it or getting lucky. Honestly I was under the impression I had already done this by spending 5-6k on a fsq. I’m actually still of the belief the fsq is likely the best scope for large sensors like we are using. But what is the alternative even if we spend countless more ? 

I agree Ken, that the FSQ has a top reputation and should be expected to be 'perfect' for that price. I was referring to several posts on the forum from people that were disappointed with the edge performance on their FSQs and ended up sending them off for realignment, to France I believe, and that they came back much improved. That was my reference to being lucky, as to whether you get a perfectly aligned one 'off the shelf'. Hopefully, someone who also uses a similar FSQ with a full frame sensor can comment as to whether what you are experiencing could be improved or not.

Alan